Appearance
Governance
Safe multisig (2-of-3) ──schedules──▶ VerityTimelock (delay) ──owns──▶ VerityCoordinator
│ ▲
└──────────────────── guardian: pause() only ─────────────────────────┘- Admin changes (fees within their caps, config, registering or removing keys, migration targets, unpausing) are scheduled on the timelock and can only execute after the delay. Anyone can watch the timelock's
CallScheduledevents to see changes coming. - Emergency pause is instant: the Safe is the coordinator's guardian and can call
pause(), which only stops new requests. Pending requests can still be fulfilled or cancelled. Unpausing goes through the timelock. - Contracts are immutable. A new version is a new deployment. Subscription owners can move their balance and consumers with
migrateto a coordinator the owner has approved (the new coordinator must also approve the old one as a source). The subscription gets a new subId there; consumer owners point their contracts at the new coordinator withsetCoordinatorand switch to the new subId. See Upgrading to a new coordinator.
| Testnet v4 | |
|---|---|
| Safe (2-of-3) | 0x555a7E7cF9f1C8f2b20D5b13a20F765678EC8d1B |
| Timelock (10 min; mainnet will use days) | 0xC6e3Aa23DDC19699bb32d8A8e82CC6644FdA099c |
The testnet coordinator is v1.0.0 and its guardian is the Safe.
Limits the owner cannot exceed
| Cap | |
|---|---|
| Premium | 100% |
| Protocol share of premium | 50% |
| Protocol flat fee | 1,000,000 PLS |
| Request timeout | at least 256 blocks |
| Confirmations | 1–200 |
A request is never charged more than what was reserved when it was made.
Upgrading to a new coordinator
The procedure (the migration path is covered by the contract and node test suites):
- Deploy the new coordinator (it can reuse the BlockhashStore). Register the same keys, then from the deployer call
setMigrationSource(old, true),setGuardian(Safe)andtransferOwnership(timelock). - The Safe schedules one timelock batch:
old.setMigrationTarget(new, true)andnew.acceptOwnership(). It executes the batch after the delay. - Each subscription owner waits until the subscription has no pending requests, then calls
old.migrate(subId, new). TheSubscriptionMigratedevent on the new coordinator gives the new subId. - Each consumer owner calls
setCoordinator(new)and switches to the new subId (DiceRoll:setRequestConfig). - Operators point their nodes at the new coordinator (
coordinatorandstart_block) once the old coordinator has no pending requests left for their keys. On restart the node notices the change, archives its records for the old coordinator (tablerequests_archive) and scans the new one fromstart_block. Its scan cursor and open requests belong to the old coordinator, so it starts clean for the new one.